Related Experiment Videos
Validation of an automatic crackle (rale) counter
R L Murphy1, E A Del Bono, F Davidson
1Pulmonary Service, Faulkner Hospital, Boston, Massachusetts.
The American Review of Respiratory Disease
|October 1, 1989
Summary
A new computer system automatically counts lung sound crackles, offering a more objective measure than subjective listening. This automated method shows feasibility for improving noninvasive cardiopulmonary diagnosis.
Area of Science:
- Pulmonary Medicine
- Biomedical Engineering
- Acoustics
Background:
- Crackles in lung sounds are vital for clinical decision-making and disease severity assessment.
- Current methods for crackle estimation, auditory or visual from time-expanded waveforms, are subjective or cumbersome.
- Objective and efficient crackle quantification is needed to improve diagnostic accuracy.
Purpose of the Study:
- To develop and validate a computer-based system for automatic crackle counting from lung sound recordings.
- To compare the performance of the automated system against human observers and time-expanded waveform analysis.
Main Methods:
- Lung sound recordings were obtained from 41 subjects using an air-coupled electret microphone.
- A computer-based system was developed for automatic crackle detection and counting.
- Interobserver agreement and correlation with visual and automated analyses were assessed.
Main Results:
- High interobserver agreement (r = 0.88) was observed for manual crackle counts.
- Automatic crackle counts showed significant correlation with time-expanded waveform analysis (r = 0.78) and physician counts (r = 0.74).
- Automated and visual methods yielded higher crackle counts per breath compared to physician observers.
Conclusions:
- Computer-based automatic crackle counting is feasible and reliable.
- This automated approach can enhance the objectivity and efficiency of noninvasive cardiopulmonary diagnosis.
- Addressing discrepancies in crackle counting criteria may further refine automated analysis.